« first day (2627 days earlier)      last day (2691 days later) » 

17:24
Is arbitrary intersection of compact sets in $\Bbb R^k$ compact? I think this because, a set in $\Bbb R^k$ is compact iff it is closed and bounded, and arbitrary intersection of closed sets is closed, and arbitrary intersection of bounded sets is bounded. Is my reasoning correct?
Yes, that's correct
@MatheiBoulomenos der Sylowmeister ist hier
@Silent can you generalize it to arbitrary topologies?
aaah! okay!
17:41
Hi chat
Any interresting maths
Salut, @Astyx.
Hi @TedShifrin
@Leaky: I don't think normal does it. You need $\Bbb Q(\alpha)$ be be in a succession of degree 2 extensions. I believe I can give you a counterexample.
hi @Mathei
@TedShifrin I'm sure normal does it
I remember the proof
Does normality give you the sequence of degree 2 extensions?
17:44
Yes, if the Galois group is $2$-group
I'm about to witness a duel between a professor and an undergrad lol
I'm far from an algebraist, @Leaky. It's just that I've thought about this briefly when I wrote my book.
@TedShifrin since the Galois group is a $2$-group, by Sylow there is a chain increasing by $2$ each
@Astyx : je suis chat =)
@mathreadler un chat
17:45
You can always refine a normal series such that the quotients are simple
@MatheiBoulomenos in this case the quotients are all Z2
Yes, that's the point
Salut @Ted
Tu aurais un/des exercice(s) de maths pour m'occuper pendant mes amphis ennuyeux ? :)
@Astyx which topic?
Pas à ce moment-ci, Astyx.
17:46
Any really
Sinon tu vas bien @Ted ? Quoi de neuf ?
I guess the point is that in general $\Bbb Q(\alpha)\subset K$ where $K/\Bbb Q$ is Galois of degree $2^m$, but that doesn't mean that $\Bbb Q(\alpha)$ is itself in the tower.
22 hours ago, by MatheiBoulomenos
Sometimes, when I'm bored (e.g. if my functional analysis prof talks about numerics of PDEs), I just pick a random number, check if there is a simple group of that order on OEIS and if not, try to prove that there is no simple group of that order
@Astyx ^
Having lots of back pains, Astyx, so I'm not in a really good mood.
@TedShifrin but $\Bbb Q(\alpha)$ is normal
$\Bbb Q (\alpha)$ is not necessarily normal
17:48
oh
I don't want to assume that. I just want to consider $\Bbb Q(\alpha)$ inside the Galois extension.
But I haven't thought about this in 5 years, so I defer to you guys.
4 hours ago, by Leaky Nun
PROPOSITION 19.7 If $K$ is a subfield of $\Bbb R$, generated by the coordinates of points in a subset $P \subseteq \Bbb R^2$, and if $\alpha$ and $\beta$ lie in a normal extension $L$ of $K$ such that $L \subseteq \Bbb R$ and $[L:K] = 2^r$ for some integer $r$, then $(\alpha,\beta)$ is constructible from $P$.
My brain isn't going to handle this.
@TedShifrin right, I misremembered the assumption
$\Bbb Q(\alpha) \subset K$ where $[K:\Bbb Q] = 2^m$
Oh, I hope your back gets better soon !
17:49
But if we have a larger field that contains $\alpha$ and is generated by constructible elements, then this is enough
Blah. Of course my chiropractor is on vacation now for a week.
as the constructible elements are closed under field operations
so $\operatorname{Gal}(K:\Bbb Q(\alpha))$ is a subgroup of $\operatorname{Gal}(K:\Bbb Q)$ (Galois correspondence)
@LeakyNun yep
@LeakyNun I think I can't because closed and bounded iff compact does not hold for arbitrary topologies. But pardon me if im wrong.
17:51
I don't think $\operatorname{Gal}(K/\Bbb Q(\alpha))$ is relevant here
@Silent so can you construct an argument from compact alone?
@LeakyNun, oh! I have to think about it!
@MatheiBoulomenos Is $\newcommand{Gal}{\operatorname{Gal}}\Gal(\Bbb Q(\alpha)/\Bbb Q)$ a subgroup of $\Gal(K/\Bbb Q)$?
@Silent or, can you come up with a counter-example?
@TedShifrin that's just Sylow and refining
you can always come up with a tower that has $\Bbb Q(\alpha)$
If $\Bbb Q(\alpha)/\Bbb Q$ is not normal, then the expression $\operatorname{Gal}(\Bbb Q(\alpha)/\Bbb Q)$ is not defined
@MatheiBoulomenos oh right
could you give me a counterexample if we disregard normality?
17:54
@LeakyNun I am really sorry to say, but I will try that tomorrow.
@Silent alright
Sure, take any irreducible polynomial with Galois group $S_4$ and adjoin a root of that
You don't even need to compute any discriminants, it's enough to show that the Galois group contains a $3$-cycle
which is easy by reduction modulo $p$
ok thanks
Hi chat
Hi @BalarkaSen
18:06
hullo
Hey @Mathei, @Astyx
What's up ?
Not much; I was off to somewhere else, and came back home right now
@MatheiBoulomenos I feel like thinking a tiny bit about Galois theory now. Can you kindly restate the equivalent definition of an extension $K/F$ being normal that you were speaking of a few days ago? (The definition I am used to is, if $\alpha \in K$ is algebraic then all the conjugates of $\alpha$ ie roots of the minimal polynomial of $\alpha$ are all in $K$).
$K/F$ is normal iff for every $F$-algebra homomorphism $\sigma: K \to \overline{F}$, we have that $\sigma(K) = K$
Aha, thanks
18:13
By some extension theorem this is equivalent to the fact that for every $\sigma \in \operatorname{Gal}(\overline{F}/F)$ we have $\sigma(K)=K$
you can also weaken the condition to $\sigma(K) \subset K$ if you want to
I have a semigeometric interpretation of this but let me thrash out some thoughts. This is very interesting.
there's more to maths than geometry
exactly why there is room for digging out geometric analogies, @Alessandro. Otherwise that would be boring; you'd be interpreting geometry by geometry.
@Rickyfox amusingly, I now find myself needing a nice smooth-step function as well
@BalarkaSen :D
18:16
Still you can't turn everything into geometry!
I guess there's a analogy with fundamental groups and (universal) covering spaces hiding somewhere, but I can't formulate it right now
Here's a question for the room, on that note.
The reason I am interested is because once upon a long time ago I kind of figured out that $\text{Gal}(\bar{k}/k)$ is basically like a fundamental group, but with the choice of basepoint being a choice of an embedding $k \hookrightarrow \bar{k}$
@Semiclassical heh, what a coincidence
Suppose you want a function $f(x)$ that interpolates between a line $y=kx$ for $x<0$ and the line $y=1$ for $x>1$.
18:18
what do (equivalence classes of) paths correspond to in that analogy?
The most obvious approach is to do use a polynomial of sufficient degree. This is enough to ensure C^k smoothness for some finite $k$.
@Alessandro If $i_1, i_2 : k \hookrightarrow \bar{k}$ are two "points", I think the paths are morphisms $\bar{k} \to \bar{k}$ such that the triangular diagram commutes.
How would one do it if one wants C^infty smoothness at the endpoints?
I spent a long time trying to figure out what homotopies should be but was unable to succeed.
I mean, the fundamental group is (under suitable coniditions) isomorphic to the group of deck transformation of the universal cover
deck transformations = field automorphisms is not that far fetched
18:20
@MatheiBoulomenos Yes, but the analogy I am proposing is different.
Hmmm, are there a lot of embeddings $k\hookrightarrow\bar{k}$ in general?
I am actually making an analogy with the paths and homotopies definition of $\pi_1$, not the $Deck(\tilde{X}/X)$ definition.
Which is why it got me excited so much when I first thought about it
@Alessandro Loads.
@AlessandroCodenotti yes
There is no such canonical choice of an embedding
oh, obviously so
I mean there's plenty even in the nice case of $\Bbb R$ and $\Bbb C$
18:22
Which is basically why it's like the basepoint of $\pi_1(X, x_0)$ when looking at $Gal(\bar{k}/k)$
@Alessandro Right.
I guess I should imitate something along the lines of $e^{-1/x}$ for $x>0$.
And the other cases are way messier so...
Also you can prove that if you change the choice of embedding of $k$ in $\bar{k}$, you get a conjugation isomorphism $Gal(\bar{k}/k) \to Gal(\bar{k}/k)$ (conjugation by an automorphism of $\bar{k}$, not fixing anything in the base, so like a path)
Much like what you get when you change basepoint of fundamental group
Your spaces are all path connected, as always
18:25
I guess you also want stuff like semilocally simply connected (and possibly Hausdorff? not sure)
@BalarkaSen Ok, I have to go for an evening driving lesson in a moment but I'll think about it
Well, to have universal cover
you don't need that to define pi_1
@Alessandro I think all of what I am saying is known to the actual algebraic geometers, though. If $X$ is a scheme people call morphisms $\text{Spec} \bar{k} \to \text{Spec} X$ as "etale points" or some shit
I meant when you try to bring out analogies to Galois theory
18:26
@MatheiBoulomenos path connectedness is enough to get $\pi_1$ independent of the basepoint I think
@AlessandroCodenotti sure
@BalarkaSen I know some of those words... like spec and nothing else :P
I meant for the general consideration
and embeddings $k \to \bar{k}$ give maps $\text{Spec} \bar{k} \to \text{Spec} k$.
(contravariance)
Wait, I don't think we mean the same thing with Spec, what is that?
18:28
Prime spectrum
comes with a topology and sheaf of rings
I mean for fields it's kind of dumb; fields have no ideals let alone prime ideals
but apparently etale theoretically it's interesting
whatever that means
yeah that was my concern, there aren't many prime ideals in fields... or ideals at all
well I mean it has (0) but yeah, it's not set theoretically interesting
I wrote a page or two about the analogy that's on my computer somewhere but it was very pretentious
@BalarkaSen Have looked at Szamuely? I haven't but I heard good things about it
18:33
I was recommended that when I was into algebra (yes, once upon a long time ago...)
I too have heard it's really good
@BalarkaSen :O
what made you change your mind?
The fact that I was not good at it :)
From my earlier write-up: If you think of the field extension $k \to K$ as a covering space $E \to X$ and an embedding $k \hookrightarrow \bar{k}$ as inclusion of a point $x_0 \hookrightarrow X$ then pushout of the diagram $K \leftarrow k \hookrightarrow \bar{k}$, which is $K \otimes_k \bar{k}$, should be analogous to the pullback $E \to X \leftarrow x_0$, which is $p^{-1}(x_0)$
$K \otimes_k \bar{k}$ is indeed a direct sum of $\bar{k}$'s, and a map $K \to K \otimes_k \bar{k}$ is then analogous to the inclusion $\sqcup y_i = p^{-1}(x_0) \to E$ of the preimage
This is so pretentious lmao
As you mention the tensor product: it's not just a $\overline{k}$ vector space, it's an algebra. And an interesting one as well: $K/k$ is separable iff $K \otimes_k \overline{k}$ is reduced
OHH
Reduced $k$-algebras are exactly the ones which are a direct product of fields right?
(I forgot the general definition)
18:50
No
it just means no non-zero nilpotents
Ive never really tried learning algebra. I'm an engineer so applications are always in focus, and often the big three calculus, probability and linear algebra take precedence. You sure can build cool stuff with abstract algebra too, but it's so... abstract!
@Mathei I made an edit. Hm, I see
If you have finite-dimensional as well (which is equivalent to Artinian and to the finiteness of the prime spectrum), then being a product of fields is equivalent to being reduced yes
Ah ok gotcha
Thanks
Algebra is cool man, I kinda wish I understood it a bit more
8
what is happening here
18:54
Hey @Eric
I think you need finitely-generated as an algebra for my parenthetical remark
Okay. I can believe that
@Balarka I was thinking about your geometric interpretation of Hilbert's basis theorem. If you have that every algebraic subset may be cut out by finitely many polynomials, that doesn't quite give you the Noetherianness $k[x_1, \dots, x_n]$. Radical ideals correspond to algebraic subsets, but you still have to worry about non-radical ideals.
But there's a lemma in commutative algebra (due to Cohen) that it is enough to know that all prime ideals are finitely generated for a ring to be Noetherian. If you translate this to geometry, this means that it is enough to know that irreducible subsets can be cut out by finitely many polynomials. From a geometric viewpoint, this is surprising as the fact that every algebraic subsets has only finitely many irreducible components is typically deduced from Noetherianness as well.
Assuming that $k$ is algebraically closed for simplicity
19:09
Ah, so you're thinking about the opposite direction (I claimed Notherianness of $k[x_1, \cdots, x_n]$ implies algebraic sets are cut out by finitely many polys). Yeah I have vaguely heard about Cohen's theorem
Can I ask a chemistry equilibrium question? (After presenting proper effort of course...)
@MatheiBoulomenos Aha, right. That is mildly surprising.
Or to put it differently, by some algebraic miracle the fact irreducible algebraic subsets may be cut out by finitely many polynomials implies that every algebraic set is a finite union of irreducibles
how on earth is Balarka talking about algebra
2
i am everywhere mwahahaha
except set theory i guess lol i dunno anything about that shit
@MatheiBoulomenos This is kinda cool.
19:23
@BalarkaSen in time you'll fall in love with it :P
@MatheiBoulomenos I am glad you are enjoying the exercise
@TobiasKildetoft what exercise?
kids these days have no mathematical chastity
spiritually degenerate bunch
@LeakyNun I assume the last one of that set
hmm
who is starring everything
19:26
@LeakyNun How to get three dots of therefore using mathjax?
hey chatterooooony
@Abcd $\therefore$ \therefore
@Abcd Don't. That symbol should really never be used
@TobiasKildetoft Why not?
19:26
@MeowMix hey dolan duck
as shorthand, it's fine.
i had a math competition
@Abcd Because there is a limit to how many symbols can be used and still end up with readable math
it was fun
here's a cool (and not too hard) question that i got:
I guess I'd say there's no reason to use it in Latex
19:27
@TobiasKildetoft Come on, it's such a frequently used symbol.
probability that the sum of the squares of two real numbers inclusively between 0 and 1 is greater than 1
if you want to use it while writing down HW quickly, I don't mind it
@Abcd Outside pure logic it really is not
@MeowMix geometric probability
19:28
yep
sum of squares = distance from origin squared
so it is 1 - pi/4
yep
#2
in general when I see two uniform distributions I think in terms of the unit square
now let's turn it into a problem in Lebesgue theory lol
@AlessandroCodenotti
yeah, have fun with that
19:29
@TobiasKildetoft Well, I am studying pre-calculus and I use it often. I have been using it since 5 years frequently. It's a commonly used math symbol in my country.
@Semiclassical hmm, I wonder what the pdf of U(0,1)^2 is
@Abcd Internationally it is not. I don't think I have ever seen it used "for real"
Oh, I see.
you see it as shorthand occasionally
I've never seen it before
19:30
what is the math jax code for SI units @LeakyNun? Like atm, L, etc.
but you'd rarely see it in anything that's not scratchwork
I mean, $\displaystyle \int_0^a f(x) \ \mathrm dx = \int \int_{\{(x,y) \mid xy < a\}} \ \mathrm dy \ \mathrm dx$
@Semiclassical I don't think I have ever encountered that outside a few students (and not the ones who write the most coherently)
@Abcd 3~\text{atm} $3~\text{atm}$
Okay @Semiclassical and @TobiasKildetoft.
@LeakyNun Cool :-)
19:31
there is a units package for SI
but if you're just doing units casually then one needs to use \text{} appropriately or just drop out of math mode momentarily
@LeakyNun has mastered MathJax.
e.g. $3$ atm
$$\begin{array}{rcl} \displaystyle \int_0^a f(x) \ \mathrm dx &=& \displaystyle \int \int_{\{(x,y) \mid xy < a\}} \ \mathrm dy \ \mathrm dx \\ &=& \displaystyle \int_0^1 \int_0^{\min(1,a/x)} \ \mathrm dy \ \mathrm dx \\ &=& \displaystyle a + \int_a^1 \frac a x \ \mathrm dx \\ &=& a - a \ln a \end{array}$$
$$f(x) = - \ln x$$
wait, this is amazing
@Abcd far from it
@Semiclassical amirite?
looks sensible
I wonder if I can generalize that result
^ is usually the beginning of disaster
19:38
lol
wait, pdf needs to be positive
wait, it is positive
smacks myself
I was talking about this physical chem question:
0
Q: Moles of a solid substance required for the attainment of equilibrium

Abcd The equilibrium constant $K_p$ at $80^{o} C$ is $1.57$ for the reaction. $\ce{PH_3BCl_3(s) <=> PH3(g) + BCl3(g)}$ . What is the minimum amount of $\ce{PH_3BCl_3}$ that must be placed in a $0.5 L$ vessel at $80^{o}C$ if equilibrium is to be attained? Since $K_p= 1.57 \therefor...

@Abcd how is R 0.0821?
@LeakyNun per litre per atm per K.
no, liter atm per K per mol
19:46
@LeakyNun tbh, i had the same thought process when I saw that f(x)
Oh yes.
@Semiclassical lol
Not bad. The reviewer uses the word "beautiful" three times in the review, as well as exciting, nice and important. And all the stuff he feels should be changed are minor things
If only he hadn't taken more than a year, I would have been completely satisfied :)
19:47
...
> The equilibrium constant $K_p$ at $80^\circ\text{C}$ is $1.57$ for the
> reaction:
>
> $$\ce{PH3BCl3(s) <=> PH3(g) + BCl3(g)}$$
>
> What is the minimum amount of $\ce{PH3BCl3}$ that must be placed in
> a $0.5~\text{L}$ vessel at $80^\circ\text{C}$ if equilibrium is to be attained?

Since $K_p = 1.57$, so $p_{\ce{PH3}} = p_{\ce{BCl3}} = 1.253~\text{atm}$

We also know that:
$$\frac{PV}{RT}=n$$
where $V=0.5~\text{atm}, P=1.253~\text{atm}, R=0.0821~\text{L}~\text{atm}~\text{K}^{-1}~\text{mol}^{-1}, T=353~\text{K}$.
that's a bit of a delay :/
@Abcd next time, I won't even look at it if I see one spelling mistake
@LeakyNun look at what?
@Abcd your latex
19:49
@Semiclassical Yeah, but at least it got accepted. It would have been a pain to find another place to submit it, as it is slightly longer than most journals will take
@LeakyNun where did I make a spelling mistake?
gotcha
any reason why it took that long?
@Abcd everywhere
@Semiclassical Not that they told us about
19:50
dissociation, uncertain @Abcd
i find myself less annoyed at the reviewer and more at the journal tbh
if it's past a certain point it seems like the journal should just choose a different reviewer
@Semiclassical My guess would be that it has been with the reviewer for most of that time, but hard to know
@LeakyNun I am not an expert typer. Anyway, I will make sure that I proof read be4 asking u for latex verification from next time.
@Abcd also, what is N2O4?
hmm, i guess you wouldn't know if they'd changed the reviewer
19:52
@LeakyNun Don't tell me that i typed that in my question.
@Abcd so I typed that there?
Hmm, it seems they have an average time from submission to acceptance of 7.7 months, so this was certainly on the long end. But it seems they take even longer to actually publish stuff (average of 20 months from acceptance to print, and almost 16 months till electronic posting)
No idea why they have such a huge backlog. It is not like they just accept everything (I assume)
huh.
that seems unusual in this day and age
@LeakyNun Edited, Previous question I solved was related to N2O4 so....
Might be due to being society owned rather than being part of one of the usual big publishers
Not sure if that has some big impact on everything I was not aware of
19:54
i guess
publishing is weird, though.
Hmm, JAMS doesn't seem to have quite as long a wait until print (and very short until electronic), but proceedings seems to be about the same
So obviously we should have submitted to JAMS instead :)
20:15
so you wouldn't have gotten as jammed up going to JAMS
neat
Does that mean we got tammed up? (is it even called TAMS?)
taco
21:19
@Astyx Bonsoir
21:52
15
Q: Evaluate $\int_0^{\infty} \frac{1-e^{-ax}}{x e^x} dx$

shimeeI found two different approaches, both is giving the same answer. Fubini: $$ \begin{align} \int_0^{\infty} \frac{1-e^{-ax}}{x e^x} \,dx &= \int_0^{\infty} e^{-x} \int_0^a e^{-xy} \,dy\, dx \\ &= \int_0^a \int_0^{\infty} e^{-x(1+y)}\, dx \,dy \\ &= \int_0^{a} \frac{1}{1+y}\, dy\\ &=\log (a+1) , a

This question has been unanswered for a fairly long time. Would be nice if someone could take the time to compose an answer.

« first day (2627 days earlier)      last day (2691 days later) »